msgBacklog, storageSize, backlogSize metrics keep increasing after stopping all consumers

**Describe the bug**
Anytime after I start a consumer for a topic and stop it when the backlog is >= 0, the msgBacklog, storageSize, backlogSize metrics continuously rise. Expected these metrics to remain constant when the consumers stop.

**To Reproduce**
Steps to reproduce the behavior:
1. Create topic to test
2. Check topic stats via `pulsar-admin topics partitioned-stats my_topic`
3. Run a consumer for `my_topic`, track the msgBacklog, storageSize, backlogSize metrics and then stop the consumer and publisher.
4. Check topic stats again. The msgBacklog, storageSize, backlogSize metrics should be increasing.

**Expected behavior**
Expecting metrics to stay constant when producer and consumer have been stopped. Even when only the consumer has existed in for that topic, the same unexpected behavior occurs

**Desktop (please complete the following information):**
- OS: MacOS 12.4

**Additional context**
Using Python 3.7.13, and Pulsar version 2.10.1(stable) standalone. Got the same behavior on previous versions on Pulsar

Reproduce the issue with a standalone Pulsar 2.10.1 topic using `pulsar-admin topics partitioned-stats my_topic`, stopping the consumer and publisher, then observe the three metrics. Trace the metric updates after consumers stop; done means msgBacklog, storageSize, and backlogSize remain constant when no producer or consumer is active.

Written by the indexing model from the issue text.